We run splyt.fit, the site for Splyt (a workout tracker for iPhone and Apple Watch). It's a Vite + React SPA on Vercel, and we wanted it crawlable without migrating to a framework.
What we did: edge functions server-render the SEO surfaces, about 2,400 pages in all. That covers 1,300+ exercise pages with instructions and demos, ready-made workouts, strength standards for every bodyweight, split guides, a help center and a powerlifting data page. vercel.json rewrites route each section to its renderer, and each renderer imports the same data module the SPA uses. So the HTML Google sees and the page a user gets come from one source, and the SPA hydrates on top.
A few things that worked well:
- Pages too thin to be useful return a 404 with noindex instead of an empty template.
- Sitemaps are split by section, so Search Console shows indexing per page type.
- The homepage is rendered by middleware, so it ships real content instead of an empty root div.
Happy to answer questions about the setup.
https://splyt.fit/exercises · https://splyt.fit/strength-standards · https://splyt.fit/data/what-lifters-total